Mistral AI vs Anthropic vs DeepSeek vs Microsoft

Head to head on 19 boards: Anthropic places best on 14, DeepSeek places best on 3, Mistral AI places best on 1 and Microsoft places best on 1. All 4 ranked on 2 of 26 boards.

Two at a time on this screen. Tap one to swap it in.

Mistral AIAnthropicDeepSeekMicrosoft

Board	Read	Mistral AI	Anthropic	DeepSeek	Microsoft
Chat2
Text ArenaArena ratingtodaytoday	Nov 9todaytoday		#41,505best of the pickedClaude Opus 5.5	#321,475DeepSeek V4.1 Flash
MultiChallengeScoretodaytoday	Nov 9todaytoday		#1067.7%Claude Haiku 5.5	#574.9%best of the pickedDeepSeek V4P1 Flash
Reasoning7
MMLU-ProAccuracy · staletodaytoday	Nov 9todaytoday	#9467.9%Mistral Large 2411	#489.1%best of the pickedClaude Opus 4.6	#3385.0%DeepSeek-V3.2	#7576.0%Phi 4 Plus
LiveBenchGlobal averagetodaytoday	Nov 9todaytoday	#5071.8%Mistral Large 4 High	#183.4%best of the pickedClaude Fable 5.1	#881.1%DeepSeek V4.1 Flash
AA IntelligenceIndextodaytoday	Nov 9todaytoday	#2726.1Mistral Large 4 High	#259.1best of the pickedClaude Opus 5.5	#1632.6DeepSeek V4.1 Flash
GPQA DiamondAccuracytodaytoday	Nov 9todaytoday	#4390.2%Mistral Large 4 0 OSS	#395.6%best of the pickedClaude Sonnet 5.5	#2193.2%DeepSeek V4.1 Flash
Epoch ECIECItodaytoday	Nov 9todaytoday	#43154.7Mistral Large 4 High	#3168.4best of the pickedClaude Opus 5.5	#19159.3DeepSeek V4.1 Flash
ARC-AGI-2Scoretodaytoday	Nov 9todaytoday	#840.0%Magistral	#393.3%best of the pickedClaude Opus 5.5	#1978.1%DeepSeek V4.1 Flash
HLEAccuracytodaytoday	Nov 9todaytoday	#404.5%Mistral 3	#246.5%best of the pickedClaude Fable 5.1
Coding4
WebDev ArenaArena ratingtodaytoday	Nov 9todaytoday	#391,560Mistral Large 4 High	#11,814best of the pickedClaude Opus 5.5	#141,665DeepSeek V4.1 Flash
SWE-bench VerifiedResolved · staletodaytoday	Nov 9todaytoday	#2256.4%Devstral Small (2512)	#176.8%best of the pickedClaude Opus 4.5	#1170.0%DeepSeek-V3.2
Aider PolyglotPass rate · staletodaytoday	Nov 9todaytoday	#4611.1%Codestral 25.01	#1072.0%Claude Opus 4	#874.2%best of the pickedDeepSeek-V3.2
SWE-Bench ProResolvedtodaytoday	Nov 9todaytoday		#198.0%Claude Opus 5
Agents and tools3
BFCLOverall accuracy · staletodaytoday	Nov 9todaytoday	#3438.4%Mistral Large 2411	#177.5%best of the pickedClaude Opus 4.5	#1256.7%DeepSeek-V3.2	#5028.8%Phi 4
MCP AtlasPass ratetodaytoday	Nov 9todaytoday		#387.2%Claude Fable 5.1
Search ArenaArena ratingtodaytoday	Nov 9todaytoday		#51,229Claude Fable 5
Vision2
Vision ArenaArena ratingtodaytoday	Nov 9todaytoday	#621,230Mistral Large 4 High	#11,309best of the pickedClaude Fable 5	#341,263DeepSeek V4.1 Flash
Document ArenaArena ratingtodaytoday	Nov 9todaytoday		#11,499Claude Fable 5
Image2
Image EditArena ratingtodaytoday	Nov 9todaytoday				#41,467MAI Image 2.7
Text-to-ImageArena ratingtodaytoday	Nov 9todaytoday				#61,330MAI Image 2.6
Speech2
Open ASRAverage WERtodaytoday	Nov 9todaytoday	#367.44 WERVoxtral-Small-24B-2507 (mistralai)			#35.55 WERbest of the pickedazure-speech-09-2026 (microsoft)
Speech ArenaArena ratingtodaytoday	Nov 9todaytoday	#151,217best of the pickedVoxtral 3 TTS			#311,145Azure Hd 2.5
Embeddings1
MTEBMean task scoretodaytoday	Nov 9todaytoday				#174.3%harrier-oss-v1-27b (microsoft)
Speed1
AA SpeedOutput speedtodaytoday	Nov 9todaytoday	#2114 tok/sMistral Large 4 High	#931 tok/sbest of the pickedClaude Sonnet 5.5	#1323 tok/sDeepSeek V4 Pro 0813
Price1
AA PriceBlended pricetodaytoday	Nov 9todaytoday		#1$0.20best of the pickedClaude Haiku 5.5	#3$0.53DeepSeek V4.1 Flash
Usage1
OpenRouterWeekly tokenstodaytoday	Nov 9todaytoday		#66.8T tokClaude Opus 5.5	#156.0T tokbest of the pickedDeepSeek V4.1 Flash

Green marks the best placing among the picked where two or more are ranked; ties are marked for each. Ranks are each board's own order (lower is better), scores as the board published them. Dash: not on that board.